Inferring Player Location in Sports Matches: Multi-Agent Spatial Imputation from Limited Observations
This work addresses the challenge of analyzing player behavior in sports analytics by making downstream tasks like physical metrics and pitch control more accessible from inexpensive event data, rather than costly optical tracking, though it is incremental as it builds on existing methods for a specific bottleneck.
The paper tackles the problem of imputing agent locations in multi-agent systems with non-uniform timesteps and limited observability (~95% missing values), specifically in football, by using LSTM and GNN components to predict player locations from sparse event data, achieving an error reduction of ~62% to within ~6.9m compared to baselines.
Understanding agent behaviour in Multi-Agent Systems (MAS) is an important problem in domains such as autonomous driving, disaster response, and sports analytics. Existing MAS problems typically use uniform timesteps with observations for all agents. In this work, we analyse the problem of agent location imputation, specifically posed in environments with non-uniform timesteps and limited agent observability (~95% missing values). Our approach uses Long Short-Term Memory and Graph Neural Network components to learn temporal and inter-agent patterns to predict the location of all agents at every timestep. We apply this to the domain of football (soccer) by imputing the location of all players in a game from sparse event data (e.g., shots and passes). Our model estimates player locations to within ~6.9m; a ~62% reduction in error from the best performing baseline. This approach facilitates downstream analysis tasks such as player physical metrics, player coverage, and team pitch control. Existing solutions to these tasks often require optical tracking data, which is expensive to obtain and only available to elite clubs. By imputing player locations from easy to obtain event data, we increase the accessibility of downstream tasks.